Original title: Mental Health Literacy Across Psychology Students and Large Language Models
Eileen Richter, Konrad Schöniger, Maya Schulze, Luisa Frede +3 · September 2026
Many people believe false ideas about mental health. Researchers wanted to know how well psychology students spot these misconceptions compared to artificial intelligence programs called large language models. They tested human students at different points in their studies along with four computer models. Everyone had to sort statements into mental health myths or true facts. The students were better at spotting false myths than identifying true facts. More experienced students scored higher than beginners on both types of statements. Overall, the computer models performed better than the students, but their accuracy changed depending on the specific model, especially on true facts. Humans and computers struggled with some of the same facts, while each side also had unique strengths on certain statements. The models still made the same errors across repeated tests. These results show where psychology training and computer tools still need work.
Imagine sorting a pile of coins into pennies and dimes. You quickly spot the pennies because they are distinct, but you often mistake some dimes for nickels. A more experienced friend does better, but still misses some dimes. In this paper, the coins are mental health statements, students sort them into myths and facts, and even a machine sorting the same coins many times made similar mistakes.
Mental health misconceptions can lead to real harm. This study shows that even psychology students—who are training to help—struggle to identify accurate facts, and that AI models, while strong, still make recurring errors. Knowing where both humans and machines fail can improve mental health education and the reliability of AI-generated information.
The abstract doesn't explicitly state limitations, but it notes that only 15 myths and 15 facts were used, and only four LLMs were tested. The findings may not generalize beyond these items and models.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Mental Health Literacy Across Psychology Students and Large Language Models — Eileen Richter, Konrad Schöniger, Maya Schulze et al. — PsyArXiv — doi:10.31234/osf.io/e6s4t_v1
Abstract licence: CC BY
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.